| OCR Text |
Show CHAPTER 1 CONCURRENT SCALABLE COMPUTING: FOUNDATION AND DEFINITIONS Although designations such as "parallel," "parallel distributed," and "distributed" are widely used and readily recognizable within almost every computing context, there is still no consensus on a precise definition describing the intended notion or style of computing being referenced. This is an inherent characteristic of a relatively new and growing technology. As such, the above designations are being used in the literature to mean various architectures and configurations ranging from tightly coupled, shared memory architectures to arbitrary configurations of machines in a local area network performing totally unrelated tasks. Often the intended meaning of these designations can only be inferred from the context of usage. Consequently, these designations have come to represent generic labels for computing environments that are characterized by some degree of concurrency and division of the work load among several processing nodes. This thesis will attempt to address these two characteristics of concurrency and division of the work load as applied to a large scientific problem. It is imperative, however, to first present a set of classifications and definitions that will lend a perspective on the discussion and results to follow. These classifications and definitions are by no means exhaustive nor are they necessarily definitive. They are only loose definitions intended as a foundation and a framework of likely options for the solution of the problem at hand. |
